Function tokenize_function

posttrain.py:133–146

(examples)

        padding = "max_length" if args.pad_to_max_length else False

        def tokenize_function(examples):
            # Remove empty lines
            examples[text_column_name] = [
                line for line in examples[text_column_name] if len(line) > 0 and not line.isspace()
            ]
            return tokenizer(
                examples[text_column_name],
                padding=padding,
                truncation=True,
                max_length=max_seq_length,
                # We use this option because DataCollatorForLanguageModeling (see below) is more efficient when it
                # receives the `special_tokens_mask`.
                return_special_tokens_mask=True,
            )

        with accelerator.main_process_first():
            tokenized_datasets = raw_datasets.map(